# File Format & Linguistic Specifications: QualiaDB Hypermedia Ecosystem > *"Pure Linked Data from Authoring to Wire: The End of XML, HTML, and JS Bloat"* **Copyright © 2026 Timothy Charles Holborn.** All rights reserved. **Principal / inventor:** Timothy Charles Holborn <timothy.holborn@gmail.com> QualiaDB defines a suite of specialized, high-performance binary, graph, and human-readable container formats optimized for decentralized knowledge representation, 3D spatial assets, LLM inference tables, universal linguistic expression, and hypermedia mindware. --- ## 1. The QualiaDB Core File Format Suite | Format | Extension | Primary Role | Underlying Architecture | | :--- | :--- | :--- | :--- | | **Q42** | `.q42` | Primary Knowledge & Phonological Graph | Merkle-CRDT DAG, Nquins (`<<[ s p o g prov ]>>`), Phonemic Nodes, Sonic Tokens | | **D10** | `.d10` | 10D Projector, Spatial Mesh & Articulatory EMF | 10D Dimensional manifold (`[α, μ, σ, t, ...]`), Vocal Tract Kinematics, 3D Mesh | | **p64** | `.p64` | Columnar Tables & Neural Acoustic Tensors | 64-aligned continuous embeddings, discrete neural audio tokens (EnCodec), Forge weights | | **HCF** | `.hcf` | Hypermedia Content Document | RDF-Native (`yaml-ld-q42` source / `CBOR-LD` binary) with Context Markup | | **HMC** | `.hmc` | HyperMedia Container Archive | Cryptographic archive packaging HCF, Q42, D10, and p64 assets with Bao streaming | --- ## 2. Universal Linguistic Support: The Ontological Grapheme Architecture Traditional computing architectures rely strictly on Unicode codepoints and standard BCP 47 language tags. As a result, thousands of indigenous, unencoded, endangered, ancient, and gestural/signed mother tongues cannot be natively digitized. QualiaDB resolves this by treating language not as flat character strings, but as **first-class semantic nodes (Ontological Graphemes)** within the Q42 graph. ### A. The Semantic Grapheme Model When a document contains text in an unencoded script, each glyph or semantic unit is represented as a structured node: ```yaml # yaml-ld-q42 representation of an unencoded project grapheme - "@id": "grapheme:project_glyph_042" "@type": "qualia:OntologicalGrapheme" "qualia:language": "@id": "lang:example_unencoded" "qualia:tag": "und-x-project" "qualia:family": "UnencodedPlaceholder" "qualia:visualGeometry": "asset://mesh/project_glyph_042.d10" # 3D/2D vector geometry "qualia:strokeSequence": "asset://strokes/glyph_042.svg" # Stroke order data "qualia:phoneticGrounding": "grapheme:project_phoneme_042" # Linked phonological node in Q42 "qualia:ipa": "ŋal" # IPA approximation "qualia:puaMapping": "\\uE042" # Unicode Private Use Area fallback "qualia:semanticConcept": "concept:FreshwaterSpring" # Direct ontological grounding ``` ### B. Multi-Modal Rendering & Accessibility 1. **Visual Display:** Poet renders the character using direct GPU mesh rendering (`.d10`) or Bezier paths, entirely bypassing system font limitations. 2. **Audio Synthesis:** Non-written or oral-first languages synthesize speech directly through the generative articulatory pipeline without text-to-speech approximations. 3. **Gestural & Signed Languages:** Link directly to 3D skeletal animation kinematics in `.d10` representing spatial signs. --- ## 3. Native Auditory Intelligence: Generative Articulatory Speech Pipeline Traditional audio files (WAV, MP3, AAC) are flat 2D air-pressure recordings over time ("acoustic exhaust"). They are immutable, difficult to manipulate semantically, and computationally disconnected from 3D visual rendering. QualiaDB replaces flat playback with a **Generative Articulatory Speech & Auditory Pipeline** operating across the **`q42` $\leftrightarrow$ `p64` $\leftrightarrow$ `d10` Triad**: ``` ┌──────────────────────────────────────────────────────────────────────────┐ │ Generative Articulatory Audio Pipeline │ │ │ │ ┌────────────────────────────────────────────────────────────────────┐ │ │ │ 1. Q42 Phonological & Intent Graph │ │ │ │ - Phonemic Nquins (IPA / Mother-Tongue Phone Nodes) │ │ │ │ - Relational Prosody (Pitch, Stress, Emotion, Cadence) │ │ │ │ - Sonic Tokens (8-byte compact event control ABI) │ │ │ └─────────────────────────────────┬──────────────────────────────────┘ │ │ │ Semantic Intent & Prosody Map │ │ ┌─────────────────────────────────▼──────────────────────────────────┐ │ │ │ 2. P64 Neural Acoustic Latents & Tensors │ │ │ │ - 64-aligned continuous embeddings / discrete EnCodec tokens │ │ │ │ - Forge WGSL kernel compute graphs & residency planner │ │ │ └─────────────────────────────────┬──────────────────────────────────┘ │ │ │ Continuous Acoustic Instructions │ │ ┌─────────────────────────────────▼──────────────────────────────────┐ │ │ │ 3. D10 Articulatory Kinematics & The 10D EMF Projector │ │ │ │ - U3 AcousticPlane (Tensor10D with shared spectral coordinate σ)│ │ │ │ - Virtual Vocal Tract (Glottal, tongue, lip kinematic manifold) │ │ │ │ - Emits physical acoustic wave & 100% synchronized 3D lip-sync │ │ │ └────────────────────────────────────────────────────────────────────┘ │ └──────────────────────────────────────────────────────────────────────────┘ ``` ### A. The Three Pipeline Pillars 1. **`q42` (The Phonological Graph):** Stores the *linguistic and phonetic intent*. Words, morphemes, and dialectal pronunciations are represented as phonemic nquins. Prosodic features (pitch contours, emphasis, emotional modulation) are stored as graph relations rather than baked into audio samples. 2. **`p64` (Neural Acoustic Tokens & Embeddings):** 64-byte aligned columnar tables store dense neural acoustic representations and ML weights. The Forge compute engine executes streaming transformations converting phonological graph sequences into continuous acoustic latents without third-party tensor bloat (no Candle/Burn runtime dependency). 3. **`d10` (Articulatory Kinematics & 10D EMF Projector):** The `.d10` container models the physical geometry and electromagnetic/spatial projection of the **virtual vocal tract**. Grounded in articulatory phonetics (glottalization, velum position, lingual resonance), the host engine projects the acoustic latents through the physical tract parameters to compute the localized sound wave. ### B. Architectural Superpowers of the Triad - **Speaker Identity Portability:** An agent’s "voice" is a `.d10` vocal tract profile. Swapping the underlying `.q42` phonological graph (e.g. from English to an example unencoded language) passes through the same `.d10` physical tract, preserving the speaker's vocal timbre and identity across all languages. - **Mathematically Perfect 3D Lip-Sync:** Because the `.d10` articulatory parameters represent the actual physical mouth/lip/jaw states used to generate the acoustics, 3D avatar facial animation and lip-sync are mathematically exact by construction. - **Unified Spectral Manifold ($\sigma$):** QualiaDB's U3 AcousticPlane shares the spectral coordinate $\sigma$ and timeline $t$ across both visual projection and auditory rendering, enabling unified cross-modal perception ("eyes and ears" intelligence). --- ## 3. Hypermedia Content Format (HCF) (`.hcf`) HCF eliminates legacy HTML, XML, and JavaScript in favor of **RDF-Native serializations**: - **Authoring Layer:** Expressed in **`yaml-ld-q42`** — clean, indentation-based semantic documents without closing tags or DOM boilerplate. - **Machine & Wire Layer:** Serialized as **CBOR-LD 1.0** — binary compressed Linked Data with shared term-dictionaries for zero-copy memory mapping on WASM, Mobile, and IoT. - **Application language:** **Vibe** (Poet), not JavaScript. Cells in HCF are Vibe (`= …`); modules are `.vibe`. JSON and JS remain a *temporary host* (Tauri IPC, Dioxus-web glue), not the format. The incremental cut from that host to this pair is in [`docs/plans/native-presentation-and-vibe-beyond-webview-2026-08-16.md`](../../plans/native-presentation-and-vibe-beyond-webview-2026-08-16.md) §0.1. HTML, CSS, JSON-LD, Turtle, Markdown, and PDF are **optional projectors** (named-profile views) of HCF/CML — doors to the traditional WWW, not the design language of this environment. The HID is not confined to those formats. The **preferred bridge** is a proportionate WASM family member plus CBOR-LD / Q42 (plan §0.3), not HTML+RDFa. COF remains a text projector for agent token windows only. CSS-as-RDF lives in the W3C archive (`http://www.w3.org/ns/css#`); it is not Qualia’s style system. See the same plan §0.2–§0.3. Do not persist an emit as the document. ### A. Document Structure in `yaml-ld-q42` ```yaml # =========================================================================== # Clinical Case Study: Comorbidity & Spatiotemporal Telemetry # Format: Hypermedia Content Format (HCF) v1.0 (yaml-ld-q42) # =========================================================================== "@context": "@vocab": "https://qualiadb.org/schema/hcf#" "snomed": "http://snomed.info/id/" "geo": "http://www.opengis.net/ont/geosparql#" "vibe": "https://qualiadb.org/schema/vibe#" "pulse": "https://qualiadb.org/schema/pulse#" "aura": "https://qualiadb.org/schema/aura#" "@id": "doc:case_study_2026_01" "@type": "HypermediaDocument" "title": "Pediatric Comorbidity Analysis" "ontology": "asset://graph/medical_ontology.q42" "content": - "@type": "Section" "heading": "Patient Assessment" "blocks": - "@type": "Paragraph" "text": "Patient presented with acute symptoms of diabetic complication." "annotations": - "@type": "ContextAnnotation" "span": [24, 45] "target": "snomed:44054006" "aura:schema": "snomed:DiabetesMellitus" "prov:confidence": 0.98 - "@type": "ReactiveCell" "@id": "cell:risk_score" "formula": "=COUNT(graph.match(?s, :hasCondition, snomed:44054006))" - "@type": "MultiModalContainer" "assets": - "@id": "asset:organ_model" "@type": "SpatialMesh" "src": "asset://mesh/pancreas_anatomy.d10" - "@id": "asset:embeddings" "@type": "TensorTable" "src": "asset://embeddings/clinical_vectors.p64" # Homoiconic VibeScript AST embedded directly in the document graph "scripts": - "@id": "script:telemetry_monitor" "@type": "vibe:PulseHandler" "vibe:onEvent": "pulse:on_pulse" "vibe:body": | on_pulse(stream) { if stream.has_alert() { pulse.broadcast("topic:alerts", "Telemetry anomaly detected"); let clinic = graph.get_node("qualia:Clinic_A"); aura.apply_schema(clinic, "snomed:EmergencyStatus"); } } ``` --- ## 4. HyperMedia Container (HMC) (`.hmc`) The `.hmc` container packages all multi-modal assets into an immutable, cryptographically verifiable archive: ``` ┌──────────────────────────────────────────────────────────────────────────┐ │ HyperMedia Container Archive (.hmc) │ │ │ │ ┌────────────────────────────────────────────────────────────────────┐ │ │ │ 1. Container Manifest & Bao Outboard Hash Tree (BLAKE3 1-KiB) │ │ │ │ - Header CID, DID Signatures, Permissive Commons Rights Scope │ │ │ └─────────────────────────────────┬──────────────────────────────────┘ │ │ │ │ │ ┌─────────────────────────────────┼──────────────────────────────────┐ │ │ │ │ │ │ │ ▼ ▼ ▼ │ │┌──────────────────────┐ ┌──────────────────────┐ ┌──────────────────────┐│ ││ HCF Document (.hcf) │ │ Q42 Graph (.q42) │ │ Spatial & Data ││ ││ (CBOR-LD binary AST) │ │ (Merkle-CRDT DAG) │ │ (.d10 mesh, .p64) ││ │└──────────────────────┘ └──────────────────────┘ └──────────────────────┘│ └──────────────────────────────────────────────────────────────────────────┘ ``` --- ## 5. Streaming & Distributed Synchronization Protocols ### A. Bao Verified Streaming (BLAKE3 Chunk Trees) To support seamless, real-time access to large multi-modal assets (`.p64` embedding tensors and `.d10` 3D meshes) packaged within `.hmc` archives: - **Zero Monolithic Downloads:** The HMC container stores an interleaved Bao outboard tree containing BLAKE3 parent hashes over 1-KiB leaf chunks. - **Random-Access Verification:** Webizen, mobile, and IoT clients can seek to arbitrary byte offsets (e.g. streaming slice `[102400..204800]`), verifying each chunk against the container's root hash in $O(\log N)$ time before executing or rendering. ### B. Merkle-CRDT State Compaction & Tombstone Pruning For `.q42` knowledge graphs subject to frequent edits and retractions across offline agents: - **Tombstone Pruning:** Operational deletion records are periodically consolidated via state-based delta snapshots. - **Epoch Hashing:** `qualia_core_db` computes a new convergent Merkle root for the active state, pruning historical operational tombstones while maintaining cryptographic provenance of the active graph.